Geospatial Data Science Internship Queries

[parvathykrishnank]

Please post your queries here regarding the Geospatial Data Science Internship at UNDP

[shoebgf0150]

What is the stipend amount per month?

[parvathykrishnank]

Please refer to the ToR in the advertisement for this information.

[drshaheed]

hello ma'am the entire work should be in python code or can we use ArcGIS for creating the same?

[Swetha-Kolluri]

Entire work should be in python code only.
For the BLOG, you may do data visualization with other tools (if required).

[Vivek0000007]

We can do data visualization in any software. Right?

[parvathykrishnank]

Please make all the code submissions in Python only (including visualization). But if you want to include visualizations in your blog, feel free to use other software.

[adiradhu17]

Respected sir/ma'am
Thank you very much for selecting me for the next round of assessment..!
I have gone through the instructions and all, but it has been asked to submit your assessment before 13th March, Sunday. Unfortunately, currently my mid semester examinations are going on so I wouldn't be able to spare time.
So would it be okay, if I submit on next Sunday i.e. on 20th March or would you provide me another deadline date.
I humbly request you to look into this and do the needful.
Thanking you,
Yours sincerely
Aditya

[adiradhu17]

Name - Aditya Dharaiya
Email ID - [email protected]

[parvathykrishnank]

In order to be fair to all the applicants, and in accordance with our procurement plan, we cannot give any extension on the deadline.

[msameeruddin]

In task 1 we are supposed to get the fire counts as per the districts. In Telangana, there are about 31 districts as per Wikipedia and other sources.

Whereas, if we follow GADM, The number of sub-divisions or districts is less.

What shall we do?

[msameeruddin]

I downloaded the data from this link (https://gadm.org/download_country.html)

I clicked on the level 3 link (from the above website) and it downloaded the gadm40_IND_3.kmz file.

Once it is finished, I loaded it using geopandas.

The file (gadm40_IND_3.kmz) has 2336 rows and 12 rows. I had to perform a spatial join with the given geojson file to basically filter only Telangana.

[msameeruddin]

@maby200 can you please send a link to get this dataset.

[maby200]

That is right, but I suggest you to download the shapefile..

[msameeruddin]

Yes, I tried that but it is downloading the outermost boundary of India. Not sure why.

[msameeruddin]

Finally, I got it after considering version 4.0.4. Thanks all.

[simhareddy677]

How to submit the taks.. could you please explain it in detail..

[msameeruddin]

[maby200]

If you get stuck with those instructions you can read the link below from that text:
How to contribute to a GitHub Project

[SHRAADHA]

We work on windows or linux?
software requirement shows that we used 18.04 ubuntu .If i use anaconda for python in window 11 any problem

[Swetha-Kolluri]

No problem, you can use Python in Windows or Linux.

[SHRAADHA]

Please anyone confirm my code is send or not

[YaodongQiu]

Do we actually need to download additional data for the task 1? or the data inside dataset is sufficient for task 1? The reason i am asking is that I can't find any relevant information about administrative boundary of Telangana except longitude and latitude which is not usable without the proper label.
Appreciated the help!

[Swetha-Kolluri]

Datasets are also available here in this folder:
https://github.com/undpindia/Data4Policy/tree/main/Geospatial%20Data%20Science%20Internship/Datasets

[YaodongQiu]

I think that is level 2 data which doesn't contain district information. we need level 3 so should I download level 3 from the website?

[Swetha-Kolluri]

Yes please use Level 3 in the GADM data as mentioned in the task description.

[YaodongQiu]

ok, thanks for answering my question.

[YaodongQiu]

For Task 2: I understand we need to use the fire dataset to predict the probrability of agricultural related fire (type=0) given all other features.

But I don't know how the ESRI LULC map can be used to propose the mythology.

Any help apprieciated!

[msameeruddin]

But in the readme file, it is said that it is an unsupervised model (target variable will not be present). You are saying that we need to predict the probability of type which clearly indicates a target variable. In that case, do we need to use classification models?

[YaodongQiu]

I think you are right about that. In fact I am also confusing about the statement that we need to predict the probability and also use unsupervised model.

[parvathykrishnank]

I have updated Task 2 description to clarify this question about the probability. Please check that.

[YaodongQiu]

Thanks for the clarification

[Vivek0000007]

I am still not clear about the data. Can anyone help?

[YaodongQiu]

for my understanding, You need to go to the website. find the level 3 data , then open and load it to the notebook. Then try to use it and the dataset from the forked repo to complete task 1.

[Vivek0000007]

Can you help in downloading?

[parvathykrishnank]

The GADM data is available here - https://gadm.org/download_country.html

You can download it here and extract the boundaries for Telangana

[msameeruddin]

The only problem here is we can extract the boundaries, but there are no adm names with which we can do spatial join. The geometry column has polygons but the labels are None.

[SUNITHAMC]

Hi,
I have these two questions.

1. For the Task 1 do we need to consider all types of fires or the fire due to crop burning only?
2. Do we need to consider both Aqua fire and Land fire or only the Terra/Land fire?

[parvathykrishnank]

For Task 1, please use all the fires as available in the dataset.

[parvathykrishnank]

Since we are receiving a few questions regarding Task 2, please find some clarifications below

While Task 1 is a qualifying task, in this task we want to evaluate your methodology and approach and not focus solely on the results.

Currently, we have the following dataset, which includes all the fire information as downloaded from the NASA Fire Information dataset for Telangana.

As this is an unsupervised model, we will not specifically be testing your code for evaluation metrics but on your approach and implementation. The goal of this is to help us identify how you approach the problem, your research/ideas on how this can be implemented. We hope you highlight this in the Medium Blog that you will be submitting along with the package. Please make your submission even if you do not have a classification model, but only ideas on how to implement this.

[Vivek0000007]

Can anyone help with installation of geopandas? I have tried everything to install it but still unable to

[Vivek0000007]

These files aren't downloading

[Vivek0000007]

I tried this earlier

[msameeruddin]

https://stackoverflow.com/a/58943939

I tried the above. pipwin can help you. I do not use anaconda.

[Vivek0000007]

I don't know why but geopandas isn't getting installed on my laptop. I have tried everywhere and done everything. Please help

[YaodongQiu]

For me conda cause error too. to solve the problem I did the following:
1.use pip to install the package on me local computer not the inside conda virtual environment.
2. select the correct interpreter on VS code.

[riazsheriff17]

Can I use R programming tool for coding, mostly I am familiar with using R?

[parvathykrishnank]

No. All the analytics have to be in Python.

[Vivek0000007]

I am not able to install geopandas, do we have any other solution?

[msameeruddin]

The only option left now is to try on Google Colab. You can easily install geopandas there.

[maby200]

Do literature references count in the 600 max words for the Medium post?

[sonof777]

The instructions for submission are a bit confusing. How do I get to gitbash on this github site to clone locally? Also, how do I create a new folder in the Geospatial Data Sciences Internship folder? I'm not too familiar with this site.

[maby200]

You can see this link which is step by step.
https://www.google.com/amp/s/www.dataschool.io/how-to-contribute-on-github/amp/

But make sure you are following the steps given from the UNDP as well:
https://akrabat.com/the-beginners-guide-to-contributing-to-a-github-project/

[sonof777]

The link is a bit confusing. I can't really make sense out of it. Are there any clearer directions or is there an alternative means of submission? This is the first time I am using github. The steps I need cleared up are: forking the project Data4Policy & clone locally, creating an upstream remote and syncing my local copy, and creating a folder inside the folder Geospatial Data Science Internship.

[Vivek0000007]

I have submitted, but I am not sure if it's in the right place. I am not familiar with the site. So please confirm.

[parvathykrishnank]

I have received your pull request

[u12ce045]

I completed the project and forked/pulled github repository and its looking in my profile. so anything remaining or there any additional step to followed.

[parvathykrishnank]

We have received your submission. Thank you!

[parvathykrishnank]

Please use the guidelines in the following link for submitting the assignments. Until the pull request is received by us the submission will not be evaluated.

https://akrabat.com/the-beginners-guide-to-contributing-to-a-github-project/

[Vivek0000007]

Can you confirm whose you have received?

[msameeruddin]

Yes that would be great.

[parvathykrishnank]

I have received both your pull request.

[Vivek0000007]

Okay thank you

[u12ce045]

can you check mine also!

[riazsheriff17]

what is the deadline time (IST)?

[msameeruddin]

I think they have specified it on the mail.

[parvathykrishnank]

Hello everyone,

The Pull Request has to be submitted before the DEADLINE - March 13th SUNDAY (11:59 PM EST - March 14th MONDAY 09:29 AM IST)

[SHRAADHA]

Mam , shradha can you check code is submitted from my side.

[SHRAADHA]

file name is undp-assesment-test

[parvathykrishnank]

Yes, we have received your pull request. Thank you.

[riazsheriff17]

I have submitted my assignment as Riaz sheriff task, pls check whether it has uploaded.

[parvathykrishnank]

Yes, we have received your pull request. Thank you.

[Anjali1808]

Please check if my assignment is uploaded.

[parvathykrishnank]

@Anjali1808 - Hello, I have not yet received your Pull Request. Please make sure you have followed all the steps outlined in the Submission Process. Thank you.

[Anjali1808]

Thank you for the information. Can you please check once again?

[parvathykrishnank]

Yes, we have received your pull request. Thank you.

[YaodongQiu]

Hello dear Mantainer:

I have submitted pull request for my task as Yaodong Qiu

Thank you for this project and I feel great for the learning experience

[parvathykrishnank]

Hello,

We have received your pull request. Thank you for the submission.

[sonof777]

I tried to work it out. I'm not sure if my submission has been sent.

[yulinhong99]

Hi all!
Because of my laptop's probelm, I used my friends laptop to uploaded the submission foler, but we don't know why it still shows my friends name. We will try to figure out.(I did not submitted late). Thank you so much!

[yulinhong99]

I uploaded my solution at 11:40 pm New York time, However, I still cannot change the name, So I keep uploading it. But it failed. Sorry for the inconvenience, but the folder's name is correct.

[sonof777]

I sent the pull request and uploaded my solution yesterday. Did it come up? I didn't get any response after I sent my query, so, I just wanted to make sure.

[u12ce045]

When will be the result declared, because project will be started in 21st March?

[riazsheriff17]

is the result declared ?

[Swetha-Kolluri]

Dear All,
Thank you for your submissions. We have received some great ideas from you all. We have already invited some of you for the next round of interviews based on:
(1) Current enrollment in a university (eligibility)
(2) Completion of both the tasks in written/technical round
(3) Quality of code & blog
The final result will be announced very soon.
Meanwhile keep exploring UNDP Careers for future opportunities
Goodluck everyone!

[Vivek0000007]

Can you send the shortlist here?

[Swetha-Kolluri]

Sorry that's not possible. All official communication will be over email. If you are shorlisted, you would have already received the email (pls check the spam folder as well).